Design and implementation of a low-cost attitude and heading nonlinear estimator

In this paper we propose a nonlinear observer (i.e. a "filter") for estimating the orientation of a flying rigid body, using measurements from low-cost inertial and magnetic sensors. It has by design a nice geometrical structure appealing from an engineering viewpoint; it is easy to tune, computationally very economic, and with guaranteed (at least local) convergence around every trajectory. Moreover it behaves sensibly in the presence of acceleration and magnetic disturbances. We illustrate its good performance on experimental comparisons with a commercial system, and demonstrate its simplicity by implementing it on a 8-bit microcontroller.